About & clinical governance
Built for the gap between knowing and doing.
Rounds is an educational clinical-reasoning simulator. Learners interview fictional patients, choose examinations and investigations, make management decisions, and receive a structured debrief. It is practice software—not a diagnostic tool and not a substitute for supervision.
The Rounds team
Clinical workflow
Aditi Maheshwari
Emergency Medicine postgraduate trainee · Chief Medical Officer
Aditi designs the clinical workflow and leads the review process. A leadership role does not automatically make every case clinician-reviewed; that claim is made only at the individual case-version level.
Product & engineering
Pratyush Gupta
Founder and developer
Pratyush builds the simulator, authored-case infrastructure, grading safeguards, and the version-bound provenance system that keeps review claims auditable.
How review works
The AI plays the patient. It does not decide what is clinically true.
Rounds separates conversational delivery from authored clinical state. This keeps a case reproducible and lets its review status follow the exact content a learner encounters.
Authored as a fixed clinical case
The diagnosis, history, examination findings, investigations, treatment expectations, and grading rubric are authored before a learner starts. The conversational patient does not invent the underlying case.
Checked for internal consistency
Automated audits test case structure, diagnostic identity, investigation consistency, answer leakage, grading contracts, and important safety pathways.
Clinically reviewed by exact version
A Clinician-reviewed label is permitted only when a named reviewer has approved the exact case-file version and allowed their review attribution to be displayed.
Review invalidates when content changes
Every approval is tied to a cryptographic file hash. If clinical content changes, the badge disappears until the updated version is reviewed again.
What our labels mean
Clinician-reviewed
A named clinician approved the exact, hash-matched case version and permitted that attribution. If evidence is missing or stale, the label fails closed and is not shown.
Clinical case preview
A spoiler-free discovery page showing the presenting complaint, specialty, difficulty, and fictional demographics. A preview does not itself assert clinician approval.
Questions or corrections?
Clinical-content concerns are reviewed against the authored case version. Include the case title or patient code and the part that needs attention.